Abstract The diversity of HIV sequences and the highly polymorphic nature of MHC-Ia (HLA-A/B/C allomorphs) lead to a highly heterogeneous display of HIV peptides across individuals and create a major obstacle for the design of HIV vaccines. While the diversity of HIV sequences can be overcome through the design of mosaic antigens or immunogens focused on conserved areas of the HIV proteome, the diversity of MHC-Ia is unavoidable. HLA-E is a non-classical MHC-Ib molecule highly conserved among humans (only 2 conserved alleles) that was originally identified for its role in immune tolerance. In 2016 two studies provided the first evidence of cytolytic immune responses against MHC-E-restricted HIV/SIV peptides. They showed broad CD8 T cell immune responses against MHC-E-restricted SIV peptides in a vaccine setting, and cytolytic NK response triggered by one HIV Gag peptide displayed by HLA-E and blocking inhibitory NKG2A receptor (the latter from collaborator Dr Barker and coauthored by the PI). An HLA-E-directed vaccine strategy would be applicable to the population regardless of the MHC-Ia diversity. However HIV peptides displayed by HLA-E in HIV-infected CD4 T cells have not been identified as the only known HIV binder was imputed by sequence homology. This R21 proposal tests the hypothesis that HIV- infected CD4 T cells display a diverse set of HLA-E-restricted HIV peptides targetable by various HLA-E restricted cytolytic immune responses. We developed mass spectrometry approaches and computational tools to identify self- and virus-derived intracellular and MHC-bound peptides from various cell types, including HIV-infected primary CD4 T cells. These HIV peptides include known HIV epitopes, peptides of non-canonical lengths, peptides derived from alternate reading frames of HIV genome, and uncovered novel T cell immune responses. The coverage of the HIV proteome by MHC-peptides in various types of infected cells is uneven, includes common hot spots of nested MHC-bound peptides, which do not correlate with the density of known HIV immune responses. This suggests that the immune responses elicited in natural HIV infection do not fully cover the HIV peptidome displayed by infected cells. Only a direct analysis of HLA-E-bound peptides will identify relevant targets for immune recognition, including peptides generated from alternate reading frames of HIV proteins. Here we propose to 1) define the HLA-E peptidome presented by HIV-infected CD4 T cells, 2) Identify HLA-E- restricted cytolytic immune responses against HIV (including CD8/4 T cells and NK cells). This proposal builds on the PI's expertise in HIV antigen processing and presentation, long-time collaborator Dr Heckerman from Microsoft Research for computational analysis of HLA-E peptides, collaborator Dr Barker for expertise in NK cells, and Dr Walker for CD8 T cells expertise and access to a large cohort of HIV+ donors.